Objectives: Study 1: To establish the validity of scores on Nutrition Screening Tool for Every Preschooler (NutriSTEP), a community-based parent-administered screening tool for assessing nutrition risk, by comparing scores to an expert rating. Study 2: To demonstrate test-retest reliability of NutriSTEP.
Subjects/methods: Study 1: Parents of 269 preschoolers (of 294 parents recruited from the community), completed the NutriSTEP questionnaire; a registered dietitian (RD) assessed the nutritional status (based on medical and nutritional history, 3 days of dietary recall and anthropometric measurements) of these preschoolers and rated their nutritional risk (1 (low) to 10 (high risk)). Receiver operating characteristic (ROC) curves were used to establish validity and determine appropriate cut points based on sensitivity and specificity. Study 2: Parents of 140 preschoolers (of 161 recruited) completed NutriSTEP on two occasions. Intraclass correlation (ICC) and kappa were used to assess reliability.
Results: Study 1: Scores on NutriSTEP and the RD rating were correlated (r=0.48, P=0.01). Area under the ROC curve for the high risk RD rating (score 8+) and the moderate risk rating (score 5+) were 81.5 and 73.8%, respectively. A moderate risk cut point of >20 and high risk cut point of >25 were identified for the NutriSTEP scores. Study 2: The NutriSTEP score was reliable between administrations (ICC=0.89, F=16.7, P<0.001). Most items on the questionnaire had adequate (kappa>0.5) or excellent (kappa>0.75) agreement.
Conclusions: The NutriSTEP questionnaire is both valid and reliable for determining nutritional risk in preschoolers.